INDEX

Explanations

references to administrative procedures or requirements involving multiple individuals or accounts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oller

-0.15

uju

-0.15

 Guil

-0.14

Ã¤nge

-0.14

ÐµÐ²Ð¾Ð³Ð¾

-0.13

 Guerr

-0.13

reur

-0.13

miyor

-0.13

 Danger

-0.13

.choices

-0.13

POSITIVE LOGITS

 separate

0.16

legg

0.15

elsen

0.15

igos

0.15

 rule

0.15

^K

0.14

 Sears

0.14

/xhtml

0.14

Ð´Ð¸Ð½Ð°

0.14

ernes

0.14

Activations Density 0.208%

references to administrative procedures or requirements involving multiple individuals or accounts

No Comments

No Known Activations